analysis-workspace

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required workflow explicitly supports fetching public provider/web content at runtime (e.g., via web-request / provider-api-request with saveToFile), and then reading that saved response text back into run-code via workspaceRead, which can include outsider-authored free text (API/web page bodies) that becomes LLM context.